{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:57YS2DO77ZDCSKSROUFNVONETM","short_pith_number":"pith:57YS2DO7","schema_version":"1.0","canonical_sha256":"eff12d0ddffe46292a51750adab9a49b2c163eaf1ce8b07a22de61d1a3914a4e","source":{"kind":"arxiv","id":"2310.07152","version":1},"attestation_state":"computed","paper":{"title":"No Privacy Left Outside: On the (In-)Security of TEE-Shielded DNN Partition for On-Device ML","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CR","authors_text":"Bingyan Liu, Chen Gong, Ding Li, Xiangqun Chen, Yao Guo, Yifeng Cai, Yuanyuan Yuan, Ziqi Zhang","submitted_at":"2023-10-11T02:54:52Z","abstract_excerpt":"On-device ML introduces new security challenges: DNN models become white-box accessible to device users. Based on white-box information, adversaries can conduct effective model stealing (MS) and membership inference attack (MIA). Using Trusted Execution Environments (TEEs) to shield on-device DNN models aims to downgrade (easy) white-box attacks to (harder) black-box attacks. However, one major shortcoming is the sharply increased latency (up to 50X). To accelerate TEE-shield DNN computation with GPUs, researchers proposed several model partition techniques. These solutions, referred to as TEE"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2310.07152","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2023-10-11T02:54:52Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"7b028e2b84a307864a89377429909a31826b8bcc3d4e3caa09f8c83f34cada47","abstract_canon_sha256":"fd43984dcc2b31ef84e32c6860290e9691a85b91d4c8cbfc615dedcb7bf2fbc3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:59:41.501047Z","signature_b64":"7OEb0eSWx3muVRS2QTq/QM9w+m3I82U2ZHY2xJl45qgkXiumN7RQPdnu4JT0xkwbSOOPW1BS2UJrQ1Ik60B3CQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"eff12d0ddffe46292a51750adab9a49b2c163eaf1ce8b07a22de61d1a3914a4e","last_reissued_at":"2026-07-05T06:59:41.500558Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:59:41.500558Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"No Privacy Left Outside: On the (In-)Security of TEE-Shielded DNN Partition for On-Device ML","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CR","authors_text":"Bingyan Liu, Chen Gong, Ding Li, Xiangqun Chen, Yao Guo, Yifeng Cai, Yuanyuan Yuan, Ziqi Zhang","submitted_at":"2023-10-11T02:54:52Z","abstract_excerpt":"On-device ML introduces new security challenges: DNN models become white-box accessible to device users. Based on white-box information, adversaries can conduct effective model stealing (MS) and membership inference attack (MIA). Using Trusted Execution Environments (TEEs) to shield on-device DNN models aims to downgrade (easy) white-box attacks to (harder) black-box attacks. However, one major shortcoming is the sharply increased latency (up to 50X). To accelerate TEE-shield DNN computation with GPUs, researchers proposed several model partition techniques. These solutions, referred to as TEE"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2310.07152","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2310.07152/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2310.07152","created_at":"2026-07-05T06:59:41.500615+00:00"},{"alias_kind":"arxiv_version","alias_value":"2310.07152v1","created_at":"2026-07-05T06:59:41.500615+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.07152","created_at":"2026-07-05T06:59:41.500615+00:00"},{"alias_kind":"pith_short_12","alias_value":"57YS2DO77ZDC","created_at":"2026-07-05T06:59:41.500615+00:00"},{"alias_kind":"pith_short_16","alias_value":"57YS2DO77ZDCSKSR","created_at":"2026-07-05T06:59:41.500615+00:00"},{"alias_kind":"pith_short_8","alias_value":"57YS2DO7","created_at":"2026-07-05T06:59:41.500615+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.29450","citing_title":"Protecting On-Device AI Inference: A Systematic Review of Attacks and Defence Mechanisms","ref_index":53,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/57YS2DO77ZDCSKSROUFNVONETM","json":"https://pith.science/pith/57YS2DO77ZDCSKSROUFNVONETM.json","graph_json":"https://pith.science/api/pith-number/57YS2DO77ZDCSKSROUFNVONETM/graph.json","events_json":"https://pith.science/api/pith-number/57YS2DO77ZDCSKSROUFNVONETM/events.json","paper":"https://pith.science/paper/57YS2DO7"},"agent_actions":{"view_html":"https://pith.science/pith/57YS2DO77ZDCSKSROUFNVONETM","download_json":"https://pith.science/pith/57YS2DO77ZDCSKSROUFNVONETM.json","view_paper":"https://pith.science/paper/57YS2DO7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2310.07152&json=true","fetch_graph":"https://pith.science/api/pith-number/57YS2DO77ZDCSKSROUFNVONETM/graph.json","fetch_events":"https://pith.science/api/pith-number/57YS2DO77ZDCSKSROUFNVONETM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/57YS2DO77ZDCSKSROUFNVONETM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/57YS2DO77ZDCSKSROUFNVONETM/action/storage_attestation","attest_author":"https://pith.science/pith/57YS2DO77ZDCSKSROUFNVONETM/action/author_attestation","sign_citation":"https://pith.science/pith/57YS2DO77ZDCSKSROUFNVONETM/action/citation_signature","submit_replication":"https://pith.science/pith/57YS2DO77ZDCSKSROUFNVONETM/action/replication_record"}},"created_at":"2026-07-05T06:59:41.500615+00:00","updated_at":"2026-07-05T06:59:41.500615+00:00"}